Why You Can't Trust Your Eyes On Video Calls Anymore

Would you realize the person on your video call was AI? Tavus Griffin can respond through voice, facial expressions and gestures. We look at its real demos, how the interaction works, the 48% study result, NVIDIA’s VideoFDB benchmark and what Griffin’s limited research preview means for access. 🔔 SUBSCRIBE TO BYTEFORWARD AI news, coding agents, model comparisons, benchmarks, and the things people are actually building: https://www.youtube.com/channel/UCGGHPOAzUaKsSc2yjsOeAFQ?sub_confirmation=1 What gave it away for you—or would it have got you? CHAPTERS 0:00 Would you spot the AI? 0:14 Meet Tavus Griffin 0:55 Gestures and Simon Says 1:52 Seeing what you show it 2:19 Coaching and conversational timing 3:17 How Griffin works 3:58 What the 48% study actually tested 4:37 NVIDIA’s conversation benchmark 5:16 Who can access Griffin? 5:46 My take SOURCES AND DEMOS Griffin research, study methods and access updates https://www.tavus.io/griffin Official launch film https://www.youtube.com/watch?v=lHw6yoyPkpo https://x.com/tavus/status/2105704169009246248 Promotion conversation https://x.com/tavus/status/2105704973036310927 Simon Says demonstration https://x.com/tavus/status/2105704311330390313 Storytelling demonstration https://x.com/tavus/status/2105704541190856704 Rubik’s Cube coaching https://x.com/tavus/status/2105704783185375402 Soldering and timing demonstration https://x.com/tavus/status/2105704914156724693 NVIDIA and David AI VideoFDB research and leaderboard https://research.nvidia.com/labs/amri/projects/video-fdb/ Original demonstrations and research visuals belong to Tavus and are discussed here in ByteForward’s commentary. This is a review of published demonstrations, not a hands-on Griffin test. Research checked October 1, 2026; availability may change. No added background music. #Tavus #Griffin #AI